NAKED STATISTICS

How to analyze “the numbers behind the news [and appreciate] the extraordinary (and growing) power of data” in today’s market-driven economy.

Wheelan (Public Policy/Dartmouth Coll.; 10½ Things No Commencement Speaker Has Ever Said, 2012 etc.) extends the scope of his 2002 best-seller, Naked Economics, to encompass the statistical know-how necessary in making informed economic and other decisions. The author provides tools to help the nonmathematical reader develop an intuitive grasp of apparently arcane topics such as the “central limit theorem,” which is used to estimate likely outcomes. Using forensic medicine as an analogy, he compares a statistician to a detective gathering information at the scene of a crime. Both are frequently involved in “building a circumstantial case based on imperfect data” and are dependent on sampling techniques. Wheelan uses a seemingly high-risk marketing campaign by Schlitz beer to illustrate the point. In 1981, the company spent $1.7 million to run a blind taste test between Schlitz and Michelob, involving 100 contestants. In fact, as Wheelan shows, it was a sure winner. While the likely outcome of a random sample would be a 50/50 split, any percentage could be framed to Schlitz's advantage. The key was in the sample. Contestants were selected on the basis of their previously expressed preference for Michelob, so that even if only 30 percent chose Schlitz, the claim that Michelob drinkers chose Schlitz was still valid. The author explains how the normal distribution works and emphasizes the importance of measuring both the mean and medium in a given study. Wheelan also explains the famous brain-teasing Monty Hall problem, which has stumped experts for years.

A delightful, informative guide to an often-intimidating subject.